qt5base-lts/tests/manual/wasm/qtwasmtestlib
Morten Sørvig ad0cb1f32d wasm: add "skip" support to qtwasmtestlib
We're not looking to skip faulty tests, but there are cases
where we would like to indicate that a test function exists
but can't run because some precondition is not met.

Pick-to: 6.4
Change-Id: Ifaaafcfa7a55beaaf56d8b25fabbe3dc2566350f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Mikołaj Boc <Mikolaj.Boc@qt.io>
2022-08-17 04:52:46 +02:00
..
qtwasmtestlib.cpp wasm: add "skip" support to qtwasmtestlib 2022-08-17 04:52:46 +02:00
qtwasmtestlib.h wasm: add "skip" support to qtwasmtestlib 2022-08-17 04:52:46 +02:00
qtwasmtestlib.js wasm: add "skip" support to qtwasmtestlib 2022-08-17 04:52:46 +02:00
README.md Enhance the qtwasmtestlib with comparison functions and status reporting 2022-07-05 21:18:30 +02:00

QtWasmTestLib - async auto tests for WebAssembly

QtWasmTestLib supports auto-test cases in the web browser. Like QTestLib, each test case is defined by a QObject subclass with one or more test functions. The test functions may be asynchronous, where they return early and then complete at some later point.

The test lib is implemented as a C++ and JavaScript library, where the test is written using C++ and a hosting html page calls JavaScript API to run the test.

Implementing a basic test case

In the test cpp file, define the test functions as private slots. All test functions must call completeTestFunction() exactly once, or will time out otherwise. Subsequent calls to completeTestFunction will be disregarded. It is advised to use QWASMSUCCESS/QWASMFAIL for reporting the test execution status and QWASMCOMPARE/QWASMVERIFY to assert on test conditions. The call can be made after the test function itself has returned.

class TestTest: public QObject
{
    Q_OBJECT
private slots:
    void timerTest() {
        QTimer::singleShot(timeout, [](){
            completeTestFunction();
         });
    }
};

Then define a main() function which calls initTestCase(). The main() function is async too, as per Emscripten default. Build the .cpp file as a normal Qt for WebAssembly app.

int main(int argc, char **argv)
{
    auto testObject = std::make_shared<TestTest>();
    initTestCase<QCoreApplication>(argc, argv, testObject);
    return 0;
}

Finally provide an html file which hosts the test runner and calls runTestCase()

<!doctype html>
<script type="text/javascript" src="qtwasmtestlib.js"></script>
<script type="text/javascript" src="test_case.js"></script>
<script>
    window.onload = async () => {
        runTestCase(document.getElementById("log"));
    };
</script>
<p>Running Foo auto test.</p>
<div id="log"></div>

Implementing a GUI test case

This is similar to implementing a basic test case, with the difference that the hosting html file provides container elements which becomes QScreens for the test code.

<!doctype html>
<script type="text/javascript" src="qtwasmtestlib.js"></script>
<script type="text/javascript" src="test_case.js"></script>
<script>
    window.onload = async () => {
        let log = document.getElementById("log")
        let containers = [document.getElementById("container")];
        runTestCase(log, containers);
    };
</script>
<p>Running Foo auto test.</p>
<div id="container"></div>
<div id="log"></div>